Post #1468631

@Njord There's probably more nuance to be had in terms of *exactly* when AI is more likely than not to break down (e.g. the more complex your code, the more likely it is to write unit tests that don't work), but I don't care *that* much about understanding the Unethical Grift Machine.